Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
Microstructured Semiconductor Neutron Detectors (MSND) and
Instrumentation
Ryan G. Fronk, Steven L. Bellinger, Luke C. Henson, Taylor R. Ochs, Michael A. Reichenberger, J. Kenneth Shultis, and Douglas S.
McGregor
Semiconductor Materials and Radiological Technologies Laboratory (SMART) Laboratory Department of Mechanical and Nuclear Engineering
Kansas State University Manhattan, KS 66506
Tim Sobering, Russell Taylor, David Huddleston
Electronics Design Laboratory
Kansas State University Manhattan, KS 66506
1 of 23
1. MSND Recap – Benefits over thin-film-coated devices.
2. Single-MSND Devices – Disposable boards, Dominoes, Wireless Devices.
3. Helium Replacement (HeRep) – Direct replacement designs.
4. Arrayed MSNDs – Pixelated and large-area neutron detectors.
5. Advanced Detector Systems – Neutron Spectrometer, 1-D Pixel Array.
6. DSMSNDs – Dual-sided Microstructured Semiconductor Neutron Detectors.
MSND Device Roadmap
Thin-Film Coated Neutron Detectors • Mass-producible, inexpensive, compact,
rugged, low-voltage operation.
• Limited Neutron Efficiency (4-5%)
MSND Introduction Microstructured Semiconductor Neutron Detectors (MSNDs) • Mass-producible, inexpensive, compact, rugged,
low-voltage operation.
• Greater Neutron Efficiency (>45%)
2 of 23
1. MSND Recap – Benefits over thin-film-coated devices.
2. Single-MSND Devices – Disposable boards, Dominoes, Wireless Devices.
3. Helium Replacement (HeRep) – Direct replacement designs.
4. Arrayed MSNDs – Pixelated and large-area neutron detectors.
5. Advanced Detector Systems – Neutron Spectrometer, 1-D Pixel Array.
6. DSMSNDs – Dual-sided Microstructured Semiconductor Neutron Detectors.
MSND Device Roadmap
Diode fabrication – 12 4cm2 MSNDs fabricated per single 4-inch (110)-oriented silicon wafer. – 56 1cm2 MSNDs per 4-inch wafer. – Up to 50 wafers can be processed simultaneously. – Good diffusion in new furnace led to leakage current measurements of
~5 nA/cm2 at -3V bias. – Current record for single, un-stacked, device: 30.1±0.5%
MSND Fabrication
3 of 23
6
Mounting MSNDs
Mounting Fabricated MSNDs • MSND size has ranged from 0.25 cm2 to 4 cm2.
• Methods for mounting MSNDs to electronics has varied, but typically a disposable intermediate board is used.
4 of 23
Domino (2014) • A standard modular device that is mass producible and houses one MSND. • Individual Dominoes can be tiled together to form meter-long strings; strings can then be grouped
together to form blades of detectors. • Powered by a 1-3V input, draws ~3mW, weighs ~9 grams. • Typically 20% intrinsic thermal neutron detection efficiency; 1:106 gamma-rejection ratio.
Domino Neutron Detection System
5 of 23
Transmitter Package
Receiver Board
Transmitter Board
Wireless Sensor Network (2007) • IEEE 802.15.4 compliant • Onboard Atmel 128 processor • Modular and expandable • 50,000 cps • 20-30m indoor range • 70-100m outdoor range
Wireless Neutron Detection System
6 of 23
1. MSND Recap – Benefits over thin-film-coated devices.
2. Single-MSND Devices – Disposable boards, Dominoes, Wireless Devices.
3. Helium Replacement (HeRep) – Direct replacement designs.
4. Arrayed MSNDs – Pixelated and large-area neutron detectors.
5. Advanced Detector Systems – Neutron Spectrometer, 1-D Pixel Array.
6. DSMSNDs – Dual-sided Microstructured Semiconductor Neutron Detectors.
MSND Device Roadmap
10
Helium Replacement HDTRA1-12-C-0004
3He Direct Replacement (HeRep Mk I and Mk II)
• Designed to directly replace a standard 3He proportional counter: – 4 atm, 2-in. diameter by 6-in. long.
• Can contain moderator inside of device. • Strips of MSNDs are arranged to eliminate streaming between strips.
7 of 23
HeRep Mk I (2011)
MSND Area 1 cm2
Num. of MSNDs 64
Total Act. Area 64 cm2
Eff. Of MSNDs 7% @ ≥500keV LLD
Voltage 12V
11
• Comparing to a 4 atm 3He tube (a 3800 ng 252Cf source was placed 2 m from the face of each detector and measured for 30 min.
• HeRep was only as good as its weakest diode; there were many diodes present, making setting the LLD difficult.
Device Relative to 3He 3He Tube: HDPE 100.0%
HeRep Mk I: HDPE ~70 %
Helium Replacement (HeRep) Mk I HDTRA1-12-C-0004
8 of 23
HeRep Mk II (2013)
MSND Area 4 cm2
Num. of MSNDs 30
Total Act. Area 120 cm2
Eff. Of MSNDs 20% @ ≥500keV LLD
Voltage 12V
Power 30mA (Resting); ~130mA (Max)
12
• Comparing to a 4 atm 3He tube (a 60 ng 252Cf source was placed 1 m from the face of each detector and measured for 30 min.
Device Count Rate (cps) Relative to 3He 3He Tube: HDPE 17.13±0.099 100.0%
He-Rep: HDPE 17.60±0.102 102.74 ± 2.65% 3He Tube: Bare 3.35±0.046 100.0%
He-Rep: Bare 3.19±0.050 95.15 ± 9.04%
HDTRA1-12-C-0004 Helium Replacement (HeRep) Mk II
9 of 23
• Gamma-ray rejection ratio: ~2-7x10-5, 14 mR/hr
1. MSND Recap – Benefits over thin-film-coated devices.
2. Single-MSND Devices – Disposable boards, Dominoes, Wireless Devices.
3. Helium Replacement (HeRep) – Direct replacement designs.
4. Arrayed MSNDs – Pixelated and large-area neutron detectors.
5. Advanced Detector Systems – Neutron Spectrometer, 1-D Pixel Array.
6. DSMSNDs – Dual-sided Microstructured Semiconductor Neutron Detectors.
MSND Device Roadmap
14
Coarse 2-D Array
2-Dimensional Array (2011) • Comprised of 25 thin-film-coated neutron detectors.
• Used for calibration of diffracted thermal neutron beam.
10 of 23
15
Arrayed Devices (2011)
Arrayed 6x6 Dual-Stacked Design (2011) • Arrays can be linked together to
function as a single larger unit. • Devices can read out individually or
sum together as a single detector. • If one unit becomes inoperable, it can
be easily replaced.
11 of 23
Panel Array Mk I (2012) • 4x4 Array Elements are tiled together to form a modular
large-area neutron detector. • Tested using a 252Cf source with an applied voltage of ±8V.
– Reported count rate of 0.2 cts/s per ng of 252Cf at 2 meters. (215 kcts Per 5 minutes)
Arrayed Devices (2012)
12 of 23
Arrayed Devices (2013)
Panel Array Mk II (2013) • Contains 480 4cm2 Dominoes, each with εth ≈15% • Total area, including moderator: 3 ft. x 3 ft. • Total Thickness: 3 in.
– HDPE Thickness: 1 in. front, 1.5 in. back (internal). • Each string is read out individually; summed via software.
13 of 23
Gen. 2 Panel Array – Testing
* Denotes testing in hallway; likely leading to artificial increase in count-rate
Distance Count Rate (s-1) Count Rate (s-1 ng-1)
1 Meter 172.48 ± 0.76 3.15 ± 0.014
2 Meters 79.57 ± 0.52 1.45 ± 0.009
5* Meters 24.60 ± 0.21 0.45 ± 0.004
10* Meters 6.098 ± 0.066 0.111 ± 0.001
Background Rm2 /
Hallway
0.811 ± 0.021 0.893 ± 0.022
N/A
Arrayed Devices (2013)
14 of 23
Gen. 2 Panel Array – Testing
Angle Percent of 2 meter measurement (%)
0° 100 %
30° 98.70 %
45° 94.52 %
60° 75.65 %
90° 55.73 %
• Angle Test – Panel Array was kept in place and rotated for each test; distance
to source was 2 meters. • Area corrections were not made.
Arrayed Devices (2013)
15 of 23
– The briefcase is powered with 12V. – The current design weighs 21 lbs. Contains 84
Dominoes at ~15% efficiency each. – Data is output via a 5V TTL pulse.
Briefcase Detector Arrayed Devices (2013)
16 of 23
Distance Count Rate (s-1) Count Rate (s-1 ng-1)
1 Meter 29.76±1.16 cps ~0.54
2 Meters 14.65±0.57 cps ~0.27
5* Meters 4.39±0.17 cps ~0.08 * Denotes testing in hallway; likely leading to artificial increase in count-rate.
Arrayed Devices (2013)
17 of 23
1. MSND Recap – Benefits over thin-film-coated devices.
2. Single-MSND Devices – Disposable boards, Dominoes, Wireless Devices.
3. Helium Replacement (HeRep) – Direct replacement designs.
4. Arrayed MSNDs – Pixelated and large-area neutron detectors.
5. Advanced Detector Systems – Neutron Spectrometer, 1-D Pixel Array.
6. DSMSNDs – Dual-sided Microstructured Semiconductor Neutron Detectors.
MSND Device Roadmap
Portable Spectrometer • Light weight (~10 lbs.) portable neutron spectrometer that is populated with 9-
11 Dominoes. • Spectrometer interrogates a source until a FOM is reduced to where
identification can be made.
• FOM is found by comparing response to onboard template matching.
• Efficiency of current design can be greatly improved by implementing more Dominoes.
Portable Neutron Spectrometer (2012)
18 of 23
• Designed For S.A.N.S. – Extremely fine resolution – 32 Channels – >10% Efficient – <106 Dead-time (sensors)
• Prototype of larger array – 32, 64, 1024 Channel – Demonstrated: SNS-ORNL
24
1-D Linear Array (2009)
19 of 23
1. MSND Recap – Benefits over thin-film-coated devices.
2. Single-MSND Devices – Disposable boards, Dominoes, Wireless Devices.
3. Helium Replacement (HeRep) – Direct replacement designs.
4. Arrayed MSNDs – Pixelated and large-area neutron detectors.
5. Advanced Detector Systems – Neutron Spectrometer, 1-D Pixel Array.
6. DSMSNDs – Dual-sided Microstructured Semiconductor Neutron Detectors.
MSND Device Roadmap
26
1
10
100
1000
10000
100000
0 25 50 75 100 125 150 175 200
Expe
rim
enta
l Cou
nts
Channel
250 um Deep Microstructure 2x Detector & 6LiF Backfilled
New Design : 10 µsNew Design BackgroundOld Design : 1 µsOld Design Background
300 keV LLD
2.73 MeV Triton
Double-Stacked MSNDs
Double-stacking MSNDs • Neutrons streaming through silicon sidewalls are made
incident on a second MSND. • Thermal neutron absorption efficiency ~93% (53% for single
device). • 42.0±0.25% intrinsic thermal neutron detection efficiency
achieved. (300 keV LLD)
• Difficult to stack properly (misalignment, off rotationally, etc.).
• Device mismatching leads to poor signal integration.
• Double the capacitance, double the leakage current.
20 of 23
Dual-Side Etched MSND Devices (DSMSNDs)
• Dual-Sided Device Characteristics – Devices are fabricated exactly as single-sided devices. – Capable of batch processing; +50 per wafer, 50 wafers per batch capable. – High detection efficiency is possible with opposing DSMSND design; >79% intrinsic detection efficiency. – Fast charge-collection is possible with some designs; < 100 ns integration time.
D.S. McGregor and R.T. Klann, patent US-6545281; allowed April 8, 2003. D.S. McGregor, R.T. Klann, patent US-7164138; allowed January 16, 2007.
Dual-Sided MSNDs
Opposing DSMSND Design Interdigitated DSMSND Design
21 of 23
DSMSND Future Improvements Dual-Sided MSND Prototype Future Improvements
• Charge-collection efficiency must be improved • Device mass-fabrication must be perfected
– Etching imperfections on front and/or backside can render the diode useless. • Device depletion is not entirely understood; depletion region may not reach backside contact. • 4 cm2 diodes must be fabricated to increase absolute sensitivity and reduce complexity.
T/W_Cell20 40 60 80 100
Off-Set Dual-Sided MSND, Straight Trench, 6-LiFCell width W_Cell (um)
0.10 18.4% 17.8% 17.3% 16.7% 16.0%0.20 35.3% 33.1% 30.9% 28.7% 26.4%0.30 51.2% 46.1% 41.2% 36.3% 31.5%0.40 65.8% 57.0% 48.3% 39.7% 32.2%0.50 79.2% 65.5% 52.1% 40.1% 32.6%0.60 76.0% 58.4% 41.3% 29.9% 23.2%0.70 74.0% 53.5% 35.1% 24.6% 18.8%0.80 63.3% 48.8% 30.7% 21.7% 17.2%0.90 31.4% 32.4% 27.6% 20.9% 16.8%
Trench depth H = 500 um
T/W_Cell4 6 8 10 12
Off-Set Dual-Sided MSND, Straight Trench, 10-BCell width W_Cell (um)
0.10 18.0% 17.6% 17.1% 16.7% 16.3%0.20 34.2% 32.5% 30.7% 29.0% 27.3%0.30 48.7% 44.8% 40.9% 37.1% 33.3%0.40 61.4% 54.6% 47.7% 41.0% 34.2%0.50 72.4% 61.8% 51.2% 40.9% 34.5%0.60 66.3% 52.3% 38.2% 29.5% 23.9%0.70 62.1% 45.7% 32.0% 24.3% 19.7%0.80 38.1% 38.4% 28.1% 21.8% 18.3%0.90 8.7% 11.9% 14.8% 16.5% 17.3%
Trench depth H = 60 um
22 of 23
Discussion Slides
D1 of 15
Neutron Conversion Materials
D2 of 15
For Si, the cross over for Compton scattering to dominate interactions above photoelectric is at approximately 60 keV. We usually set the lower level discriminator at or above 5 times this value (> 300 keV) to reduce gamma ray background. Photoelectrons or Compton electrons with energies above 65 keV have transit lengths in Si >40 microns, a dimension larger than the lateral dimensions of the 6LiF filled trench devices!
Silicon Photon Cross-Section
D3 of 15
Thin-Film Coated Neutron Detectors • Neutron-converter material converts neutrons into charged
reaction products. • Mass-producible, inexpensive, compact, rugged, low-voltage
operation. • Poor neutron absorption efficiency (<15%). • Poor charged-particle counting efficiency.
• Limited Neutron Efficiency (4-5%)
D4 of 15
Thin-Film-Coated Solid State
Microstructured Semiconductor Neutron Detectors (MSNDs) • Mass-producible, inexpensive, compact, rugged, low-voltage operation. • Better neutron absorption efficiency (>52%). • Better charged-particle counting efficiency.
• Greater Neutron Efficiency (>45%)
Microstructured Semiconductor Neutron Detectors
D5 of 15
Energy Deposition • Increased likelihood of energy deposition by reaction
products; increases signal-to-noise ratio.
Neutron Absorption • Increased neutron absorption increases count rate
and therefore detection efficiency.
• Benefits – Better Uniformity Across Large Wafers
– This Leads to Uniform Responses From Each Device in an Array!
– Batch Wafer Processing (No Limit!) – Less Mechanical Damage than ICP RIE
• 3 Different perforation designs – Straight Trench – Chevron Trench – Rhombus Hole/Pillar
Anisotropic Chemical (KOH) Wet Etching of (110) Si
KOH Wet Etching
D6 of 15
Nano-6LiF Backfilling
D7 of 15
37
DSMSND Present Performance
Dual-Sided MSND Prototype Characterization
• Devices are tested for pass/fail prior to 6LiF backfilling based on diode characteristics – Leakage current vs. voltage (IV) curves are measured; < 5 nA cm-2 at -3 V bias. – Capacitance vs. voltage (CV) curves are measured. < 95 pF at -3 V bias.
IV-Testing • LC leads to high noise levels; difficulty
in resolving signal. • LC < 50 nA cm-2 is typically acceptable.
CV-Testing • Capacitance leads to weaker output
signal as well as poor pre-amp coupling. • Capacitance < 150 pF is acceptable.
D8 of 15
Sidewall Width 10 um 12 um 14 um 16 um 18 um 20 umTrench Width 30 um 28 um 26 um 24 um 22 um 20 um
Total Eff. 36.33% 35.29% 34.05% 32.61% 30.98% 29.19%0.3 MeV LLD 34.04% 33.27% 32.27% 31.09% 29.66% 28.07%0.5 MeV LLD 32.29% 31.94% 31.13% 30.12% 28.82% 27.36%
38
HDTRA1-12-C-0004 Modeling
D9 of 15
39
MSND Incident Angle
neutron converter material
semiconductorvolume
uniform parallel neutron beam
MSND Angular Efficiency Comparisons
D10 of 15
DSMSND Angular Efficiency Comparisons
DSMSND Incident Angle
D11 of 15
41
MSND Characterization HDTRA1-12-C-0004
Detector
MSND Neutron Testing • Diffracted thermal neutron beam at KSU TRIGA Mark II Nuclear
Reactor • Reactor Power – 0 to 500 kW • Thermal (0.0253eV) Neutron Flux: 1.72 x 102 {n cm-2 s-1 kW-1} • Calibrated against 3He-Gas Detector
MSND Gamma-ray Sensitivity • 137Cs source
• γ-ray Energy: 662 keV • 1 meter from MSND • Assay: 68.27 mCi • Exposure: 21.8 mR hr
-1
• 0.08 γ-ray µs-1 (per 4-cm2 area)
D12 of 15
42
MSND Device Characterization Neutron Efficiency of 4-cm2 MSND Detector
neutron converter material
semiconductorvolume
uniform parallel neutron beam
• 4 cm2 MSND, 440-µm deep trenches, 10-µs charge integration time. • 30.1 ± 0.5% at a 650 keV LLD with normal beam incidence. • 37.6 ± 0.7% at a 650 keV LLD with 45 deg. beam incidence.
D13 of 15
Stacked Perforation Design4: 10 µs preAmp integration time Pulse height spectrum taken with a 6LiF-filled microstructured semiconductor neutron detector formed from stacked 1cm2 devices. 42.0 ± 0.25% at a 300 keV LLD.
1
10
100
1000
10000
100000
0 25 50 75 100 125 150 175 200
Exp
erim
enta
l Cou
nts
Channel
250 um Deep Microstructure 2x Detector & 6LiF Backfilled
New Design : 10 µsNew Design BackgroundOld Design : 1 µsOld Design Background
300 keV LLD
2.73 MeV Triton
• Broken/Weak Wire-bonds – Stresses on wire-bonds led to intermittent noise issues. – Could induce problem with heater gun or squeezing device near bond pad.
• Cause: Improper wire-bonding technique.
Gen. 2 Panel Array – Issues Noise Issues
D14 of 15
Gen. 2 Panel Array – Issues Noise Issues
• Solder Pins – Pins were soldered onto DDBs using
low-cost lead-free solder. – Heating the pin while soldering the
DDB to the Domino melted the low-cost solder, lifting the DDB from the board.
– Pins and low-cost solder were removed and replaced with a ‘solder bump’.
• Cause: Poor materials quality.
D15 of 15